Free credit available Limited time
Available in console Data

Deploy Jupyter + PyTorch
in one click.

Notebooks with PyTorch, torchvision and torchaudio for machine learning on the CPU, on a server we run for you.

Launch Jupyter + PyTorch Browse all 48
Recommended from 8 GB RAM, from ~$0.03/hr (EU) 6 server locations No separate app fee

+$5.00 added at signup · exact template preserved · no card

What launches

The working stack,
preconfigured.

Choose the template and a project server. Selfhost.dev provisions the host and launches the application stack as one resource.

  • JupyterLab
  • PyTorch, torchvision and torchaudio
  • numpy, pandas and scikit-learn
  • Persistent home folder

Good fit for

Use it where
it does real work.

  • 01 Deep learning notebooks
  • 02 PyTorch model prototyping
  • 03 Teaching machine learning

From choice to app

A short path to
a usable deployment.

  1. 01

    Choose Jupyter + PyTorch

    The page opens the exact console template. Pick a server size and one of six locations, then submit.

  2. 02

    Set access

    There is no username. Paste the password you set at deploy time into the Password or token box on the sign-in page.

  3. 03

    Let it start

    The console tracks provisioning, gives the app a generated HTTPS address and shows host CPU and memory after launch.

  4. 04

    Make it yours

    Accept the no-mining agreement and deploy. On an organisation that has not topped up yet, a platform admin approves each new one before it starts; you get an email when it is approved or declined. Paste your password into the Password or token box. Start a notebook. PyTorch, torchvision and torchaudio are installed beside the scientific Python stack and run on the CPU.

Know before deploy

The operational details
that matter.

This is JupyterLab with PyTorch, torchvision and torchaudio added to the scientific Python stack, so the image is larger and the first boot takes longer. Project servers have no GPU, so PyTorch runs on the CPU. Your home folder persists across restarts, including notebooks, saved models and packages you add with pip install --user. Deploying it asks you to accept a no-mining agreement. On an organisation that has not topped up yet, a new one waits for a platform admin's approval before it starts.

Capacity
Recommended from 8 GB RAM, from ~$0.03/hr (EU)
Topology
One project server
Initial address
Generated HTTPS URL
Protection
Optional whole-VM backup

Also in the catalog

Related one-click apps.

Browse all 48
Data

Jupyter + TensorFlow

Notebooks with TensorFlow and Keras for machine learning on the CPU, on a server we run for you.
Data

JupyterLab

Python notebooks with numpy, pandas and scikit-learn ready to import, on a server we run for you.
Data

Qdrant

A vector database for embeddings, semantic search and RAG, on a server we run for you.

Frequently asked questions

Is Jupyter + PyTorch available as a one-click template?
Yes. Jupyter + PyTorch is one of 48 configurations in the console's global template chooser. The marketing page links directly to that configuration.
How much does it cost to run Jupyter + PyTorch?
Recommended from 8 GB RAM, from ~$0.03/hr (EU). The app adds no separate fee. You pay only for the project server it runs on, by the hour against prepaid credits.
Does Jupyter + PyTorch use a GPU?
No. Project servers have no GPU, so this template uses the CPU-only image and every model trains on the processor. Size the server for the data and models you plan to load.
How is Jupyter + PyTorch different from JupyterLab?
It is the same JupyterLab with PyTorch, torchvision and torchaudio added to the scientific Python stack. The image is larger, so the console budgets 20 minutes for the first boot against 15 for JupyterLab.
Why might a new Jupyter + PyTorch wait for approval?
On an organisation that has not topped up yet, a platform admin reviews each new Jupyter + PyTorch before it starts; its name is reserved meanwhile and you get an email when it is approved or declined. After a top-up it starts straight away. Crypto mining is banned on every server and deploying this template asks you to accept that first. To withdraw a request, delete the service.
How do I sign in to Jupyter + PyTorch?
There is no username. Paste the password you set at deploy time into the Password or token box on the sign-in page.
Can I use my own domain for Jupyter + PyTorch?
The app first receives a generated HTTPS address. After it is running, add one customer-owned domain from the resource Settings screen and follow the DNS instructions.
What does Selfhost.dev manage for this deployment?
Selfhost.dev provisions and maintains the project server, starts the Jupyter + PyTorch template and shows server CPU and memory metrics. The application and its data run on that single server; optional backups protect the whole project VM.

Exact template selected

Put Jupyter + PyTorch
on your own server.

The app adds no separate fee. You pay only for the project server it runs on, by the hour against prepaid credits. The console keeps this template selected through sign-up and onboarding.

Launch Jupyter + PyTorch